Beyond main effects assumption in conjoint analysis comparison of conjoint value analysis vs. choice-based conjoint. Statistical approach and construction of desings applied to new product development

摘要

The assumption of only main effects in Conjoint Analysis methods has created a debate whether to focus or not on the impact of interactions that were tested in this research.Different conjoint methodologies have been used in new product development studies and diverse results have come to which of them perform better to obtain information from respondents. A comparison of Conjoint Value Analysis CVA and Choice-Based Conjoint CBC surveys were undertaken to contrast them through utility scores, importance values of attributes and goodness-of-fit found in both methodologies using ready to drink beverages as the subject. The main effects assumption in the CVA composition rule was compared to the interaction terms in the CBC one. Two scenarios were developed;the first one considered inner characteristics of the subject and a sample size of 250 respondents. The second one considered the presentation characteristics of the subject and a sample size of 150 respondents. The attribute importance order in each methodology for Scenario 1 was different, while in Scenario 2 it was the same. Significant differences were encountered among attribute levels and within them; the same pattern was found in bothmethodologies and in both scenarios.The two higher total utility scores were obtained in the CBC using an interactive composition rule that considered beer, whereas in the CVA this level reported a negativeutility. In Scenario 1 a higher goodness-of-fit was found in the CBC, including significant interactions, in contrast with Scenario 2, where no interactions were found, and CVA had a higher goodness-of-fit.Despite the fact that the design and calculations of a Choice-Based Conjoint were not as straightforward and well-known as the design and calculations of a traditional conjoint analysis, the inclusion of interactions is valuable information that researchers in future studies should consider.